A SYSTEM THAT ENABLES THE DETECTION OF SENSITIVE DATA.

An AI-driven system for analyzing software logs addresses the limitations of static filters and manual review by providing real-time personal data detection and prevention, enhancing data security and compliance.

TR202417117A3Pending Publication Date: 2026-10-21TURKCELL TEKNOLOJI ARASTIRMA & GELISTIRME AS
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
TR202417117
Authority / Receiving Office
TR · TR
Patent Type
Applications
Current Assignee / Owner
Filing Date
2024-11-28
Publication Date
2026-10-21
Patent Text Reader

Abstract

This invention relates to a system (1) that enables the detection of personal data contained in log files generated in software applications and prevents the leakage of such data.
Need to check novelty before this filing date? Find Prior Art

Description

1 TARIFF A SYSTEM THAT ENABLES THE DETECTION OF SENSITIVE DATA. Technical Area This invention relates to the data contained within log files generated by software applications. identification of personal data and prevention of data leakage It is related to a system that provides. Previous Tech 10 Today, software logs are used to monitor application performance and troubleshoot problems. They are created for this purpose. However, these logs sometimes contain personal data of the user. It could compromise privacy. Commonly used in current systems. One method is to sort log files by specific keywords or data types. This involves examining the data using static filters. These filters can identify personal data. fixed patterns (e.g., email addresses, phone numbers, credit card numbers) (like this) Static filters only work with predefined fixed patterns. It detects them. However, personal data can be in a wide variety of formats and do not follow fixed patterns. There is a high probability of missing data outside of this. Dynamic data types or the new 20 There is no flexibility in recognizing the emerging data formats. Different The constant changes in data structures across sectors and applications render these filters ineffective. This leads to incorrect pronunciation of certain word or number combinations. There is also a high probability that this will be perceived as personal data, which can lead to unnecessary alerts. and can lead to erroneous interventions. 25 Another common method is to manually examine the software logs. System Managers / supervisors manually review the logs upon complaint and It attempts to find potential personal data leaks. The logs are being reviewed manually. It is very time-consuming to analyze, especially with large volumes of data. 30 Its applicability is quite low. The risk of human error is high. Manual reviews. 2 During this process, there is a possibility that personal data may be overlooked or misinterpreted. This is the issue. In large-scale systems, manual review requires a significant amount of work. This requires power and cost, which makes it unsustainable, especially in large-scale projects. income. Additionally, some systems anonymize certain data types before writing them to the logs. It either retrieves or masks the data. This method prevents personal data from being found in the logs. Masked logs don't always provide access to the necessary information. Troubleshooting. Some critical data required in these processes cannot be used due to anonymization. This can happen. Anonymization processes are still based on fixed rules and are dynamic. 10 It lacks the ability to identify personal data. This situation, as previously... It can exclude undefined or unexpected data formats. In some systems, it is also possible to retrospectively determine whether personal data is present in the logs. It is controlled through audits. With this method, 15 after the logging process is completed. The data is examined. This method does not prevent personal data from being written to the logs, it only... It provides an opportunity to intervene after the event has occurred. Therefore, real-time data It does not have the ability to stop leaks. The retrospective review process, It requires a lot of time and resources, especially when it requires an emergency response. In these situations, such delays can lead to serious consequences. 20 Therefore, considering the studies and shortcomings in the current technique... When this is the case, personal data in log files can be analyzed by artificial intelligence and machine learning. a system that enables automatic detection using learning techniques It is understood that it is needed. 25 Japanese patent number JP2016099722, which is included in the prior art. The document mentions a personal information leakage monitoring system. the purpose is to detect the leakage of personal information in the invention that is the subject of the patent document and Personal information leakage tracking 30 to determine the date and time the leak occurred A system has been developed. The system generates personally confidential information at specific intervals. 3 personal information creation section and this information is allocated with each internet connection. a database record that records data to a database tracked using the obtained IP address It includes the recorded personal confidential information and the recording date and time. There is a log history storage section that stores this information. Additionally, An external 5 that receives links to the contact address included in personal confidential information. The communication acquisition section and the database leak based on these connections. There is a leak detection section that identifies leaks. Finally, in the log history... by using the information, determine the date and time when the personal information was leaked. The system includes a section for specifying the date and time of the leak. This system contains personal information. by effectively detecting leaks and the timing of these leaks 10 It contributes to preventing security vulnerabilities. Brief Description of the Invention The purpose of this invention is to identify 15 files within log files generated by software applications. the identification of personal data contained therein and the prevention of the leakage of this data The goal is to create a system that prevents this. Another purpose of this invention is to ensure that logs generated in software applications are accurate. By analyzing data over time, potential personal data leaks can be identified. 20 and ensures data privacy by preventing the logging of this data. The goal is to implement the system. Detailed Description of the Invention The "Detection of Sensitive Data" process was carried out to achieve the purpose of this invention. A system enabling this is shown in the attached figure; Figure 1 shows a schematic view of the system that is the subject of the invention. 4 The parts shown in the figure are individually numbered, and the corresponding numbers correspond to these numbers. given below: 1. System 2. Electronic device 5 3. Data collection application 4. Server A. Log server Identifying personal data contained within log files and processing this data within 10 The system in question, which enables the prevention of leakage (1); - exchanging data using any remote communication protocol to perform and automatically process data and produce meaningful results. at least one configured to run applications and / or software that generates electronic device (2), 15 - carried out on electronic devices (2) at specific intervals and / or events log records that provide logs on any system and / or client when triggered at least one data server configured to retrieve log records from server (A) collection application (3), − using any remote communication protocol with electronic device (2) 20 the practice of establishing communication and collecting data through this communication (3) to exchange data, to automatically process data on it, and running at least one AI software program that interprets the information, manually and / or Data collection application by being triggered automatically at certain periods (3) to retrieve the data obtained and preprocess the data log 25 Preparing files for analysis, using machine learning models to analyze the processed data and in the log files in question to determine if there is any personal data, and to identify the identified personal data. to prevent the recording of files containing data and the analysis performed A 30-page document containing which personal data is located in which log file and where. The system is structured to generate reports and forward them to the relevant units. It contains a small server (4). The electronic device (2) in the system (1) which is the subject of the invention, on which the data is automatically 5. To run applications and / or software that function and produce meaningful results. smartphones, tablet computers, desktop computers or configured accordingly It is a device in the form of a portable computer. The electronic device (2) can be used remotely. to communicate with the server (4) using the communication protocol and the established communication protocol It is structured to facilitate data exchange via communication channels. The invention includes (1) data collection application (3) electronic device (2) is being carried out manually and / or automatically at certain intervals. By being triggered, access to the log record server (A) where the log records are located and to receive the requested log files and send them to the server (4) for analysis. It is being structured. 15 Communication between the server (4) and the electronic device (2) located in the system (1) which is the subject of the invention to establish and to conduct communication on the electronic device (2) through this established communication to exchange data with the data collection application (3) It is configured. The server (4) automatically processes and analyzes the data. at least one artificial intelligence that performs, produces meaningful results, and improves itself through learning. The software is configured to run. Server (4) data collection to process the log files obtained by the application (3) and log information in the data in the form of unnecessary data, blank lines, or system messages filtering, cleaning data by removing unnecessary information and the possibility of personal data being lost. 25 by identifying data fields with high frequency (email address, phone number) It is configured to be ready for analysis. Server (4) pre-processing By performing a normalization process on the transmitted data, the data is brought to the same scale. and / or to include within a range and delete data of different sizes It is configured. The server (4) extracts features from the log dataset and the raw 30 Selecting the most meaningful and useful information from the data to prepare it for analysis. 6 It is configured to analyze the log data of the server (4). to train the learning algorithm and enable the algorithm to learn meaningful information It is configured to provide. The server (4) pre-processed and The cleaned data is used as training data and input to the machine learning algorithm. and the algorithm will identify relationships and patterns in the data through trial and error. 5 It is structured to enable learning. It runs on the server (4). the calculation of errors during the learning process of the machine learning algorithm and to reduce the calculated error using optimization algorithms The server (4) is configured to process the data for each data log data set. During the processing process, we work with the training data and update the data weight by 10 It is configured to ensure that the error rate is reduced. Server (4) the machine learning model working on it validates the dataset during learning using to monitor the model's performance and the training (learning) process Once completed, the model is evaluated on test data, and the test data, Because the model consists of data it has not seen before, the model's generalization is 15. It is structured to measure ability. Server (4) test result of machine learning model a certain threshold value If this is the case, the model can be used in log record data analysis and 20 real-time log data to be analyzed by providing it as input to the model. is structured. Server (4) classification (category in the analysis model (estimation), regression (continuous value estimation), clustering (grouping), linear Analysis and estimation using regression, logistic regression, and Native Bayes methods. to analyze log records using any of the algorithms and log It is configured to identify personal data on the records. Server 25 (4) which personal data related to the analysis is in which log file and where to create a report containing this and send it via text message or email It is structured to transmit information to the relevant units via communication channels. Industrial Application of the Invention 30 7 Log produced in software applications with the system (1) which is the subject of the invention Identifying the personal data contained within the files and this data Leakage is prevented. The system in question (1) provides the user with information containing data privacy principles and 5 consent is given within the scope of the Personal Data Protection Law (KVKK). It operates. Around these fundamental concepts, the invention focuses on "Detection of Sensitive Data". It is possible to develop a wide variety of applications related to "A System Providing (1)" 10 and the invention cannot be limited to the examples described here, but mainly to the claims as stated.

Claims

8 REQUESTS 1. Identifying and storing personal data contained within log files. ensuring that data leaks are prevented; − data exchange using any remote communication protocol 5 to perform and automatically process data and make it meaningful to run applications and / or software that produce results at least one electronic device configured (2), - carried out on electronic devices (2) at specific intervals and / or Logs on any system and / or client are triggered by events. to retrieve log records from the log record server (A) including at least one structured data collection application (3), - using any remote communication protocol electronic device (2) to communicate and collect data through this communication. To exchange data with the application (3), to exchange data on 15 at least one artificial intelligence software that processes and interprets automatically to operate, manually and / or automatically at specific intervals data obtained by triggered data collection application (3) to collect and preprocess data for analysis of log files preparing, pre-processing 20 using machine learning models to analyze the data passed through and to identify any in the log files in question to determine whether it is personal data, and to identify the personal data that has been identified. to prevent the recording of the files containing them and the analysis performed which personal data is in which log file and where 25 a system characterized by having at least one server (4) configured (1).

2. It automatically processes data and produces meaningful results. a smartphone configured to run applications and / or software, 30 in the form of tablet computers, desktop computers or portable computers 9 as in Claim 1, characterized by an electronic device (2), which is a device a system (1).

3. Communicate with the server (4) using any remote communication protocol. to establish and exchange data through this established communication 5 Claim 1 is characterized by an electronic device (2) configured to be or A system like the one in 2 (1).

4. The electronic device (2) is operated manually and / or 10 that are automatically triggered at specific intervals and contain log records To access the log server (A) and retrieve the requested log files for analysis. Data collection configured to transmit to the server (4) for processing any of the above claims characterized by its application (3) a system like one of them (1).

5. To communicate with the electronic device (2) and through this communication data collection application (3) carried out on electronic device (2) characterized by the server (4) configured to carry out the transaction. a system like any of the above-mentioned requests (1).

6. Automatically processes and analyzes data, yielding meaningful results. at least one artificial intelligence software that generates and improves itself through learning. The above is characterized by the server (4) configured to run a system like any of the requests (1).

7. Log files obtained by the data collection application (3) processing and removing unnecessary data, empty rows, or system errors from log data. Filtering information in the form of messages cleans the data, removing unnecessary elements. extracting information and identifying data fields that are highly likely to contain personal data. Server 30 configured to identify and prepare data for analysis (4) like any of the above-mentioned claims characterized by system (1).

8. By performing a normalization operation on the pre-processed data. to include data within the same scale and / or range and 5 of different sizes characterized by the server (4) which is configured to delete data a system like any of the above requests (1).

9. Extracting features from a log dataset and identifying the most meaningful elements from the raw data.

10. Selecting useful information to prepare the data for analysis. from the above requests characterized by the configured server (4) a system like any other (1).

10. To train a machine learning algorithm that will analyze log data and 15 designed to enable the algorithm to learn meaningful information any of the above requests characterized by the server (4) such a system (1).

11. The pre-processed and cleaned data is used as training data by the machine. to provide input to the learning algorithm and the algorithm's trial 20 to enable it to learn relationships and patterns in the data through trial and error from the above requests characterized by the configured server (4) a system like any other (1).

12. The machine learning algorithm running on it has 25 errors during the learning process. calculations are to be performed and optimization algorithms are to be used structured to process data in order to reduce the calculated error any of the above requests characterized by the server (4) such a system (1). 11 13. For the log dataset, work on the training data in each data processing step. and to reduce the error rate by updating the data weight from the above requests characterized by the configured server (4) a system like any other (1).

14. Validation during the learning process of the machine learning model being worked on. monitoring the model's performance using the dataset and the training process Once completed, evaluate and test the model on test data. Because the data consists of information that the model has not seen before, the model 10 characterized by server (4) configured to measure generalization ability a system like any of the above-mentioned requests (1).

15. The test result of the machine learning model exceeds a certain threshold value. If it is present, the model can be used in log record data analysis. and analyze real-time log data by providing it as input to the model 15 the above requests are characterized by the server (4) configured to be so. a system like any other (1).

16. The analysis model includes classification, regression, clustering, and linear regression. Logistic regression, native Bayesian analysis and estimation 20 to analyze log records using any of the algorithms and structured to identify personal data on log records any of the above requests characterized by the server (4) such a system (1).

17. Which personal data related to the analysis is located in which log file and where? to generate a report including this and send it via text message or email structured to transmit to the relevant units through a communication channel in this form any of the above requests characterized by the server (4) such a system (1). 30